Abstract

The goal of the global path planning task is to produce an optimal safe path for robot respect to the given map. There are many global path planning methods that have been studied. In this paper, we propose a novel Hybrid Genetic Algorithm (HGA) that is used to generating smooth paths for differential wheeled robots. The main idea of HGA is to provide the dynamic mutation rate and switchable global-local search method to the mutation operator of ordinary genetic algorithm. By deploying these modifications, the premature convergence of the generic genetic algorithm and the high time-consuming fitness calculation of the memetic algorithm are reduced. HGA also takes care of chromosome length (defined by the size of a set of points that construct a path) by applying the population replacement method. Our algorithm satisfies the important criteria in the path planning task: safe and minimum-length. Based on continuous-curvature Piecewise Cubic Bezier Curve, HGA directly provides the smoothed paths for differential wheeled robots. Therefore, our proposed algorithm does not need any third-party algorithm for smoothing planned path. Several experiments regarding the proposed algorithm on our robot and its results are analyzed and presented.

摘要

全局路径规划任务的目标是为机器人根据给定地图生成最佳安全路径。已经研究了许多全局路径规划方法。在本文中，我们提出了一种新颖的混合遗传算法 (HGA)，用于为差动轮式机器人生成平滑路径。HGA 的主要思想是为普通遗传算法的变异算子提供动态的变异率和可切换的全局-局部搜索方法。通过部署这些修改，减少了通用遗传算法的早熟收敛和模因算法的高耗时适应度计算。HGA 还通过应用种群替换方法来处理染色体长度（由构成路径的一组点的大小定义）。我们的算法满足路径规划任务中的重要标准：安全和最小长度。HGA 基于连续曲率分段三次贝塞尔曲线，直接为差动轮式机器人提供平滑路径。因此，我们提出的算法不需要任何第三方算法来平滑规划路径。分析和介绍了关于在我们的机器人上提出的算法及其结果的几个实验。